RDP 9811: Effective Real Exchange Rates and Irrelevant Nominal Exchange-Rate Regimes 3. Data

Our data set consisted of 90 countries with monthly data on the effective RER, nominal exchange-rate regimes and inflation rates over the period 1978 to 1994.[14] Also, we used annual data on real GDP to calculate the variance of real GDP growth rates.

We used data from Goldfajn and Valdes (1996) on monthly effective RERs, available from January 1960 to December 1994. The effective RER is a trade-weighted average of bilateral RERs with those trading partners encompassing 4 per cent or more of trade in either imports or exports. In the construction of these bilateral RERs, the Wholesale Price Index was used if available; otherwise the Consumer Price Index was used.

We built our data set on nominal exchange-rate regimes from the monthly issues (October 1978 to November 1996) of the International Monetary Fund's International Financial Statistics (IFS) and from the Annual Report on Exchange Arrangements and Exchange Restrictions also of the IMF (1978 to 1983). The raw data on regimes describes over 25 possible exchange rate arrangements. We simplified this classification scheme by aggregating categories into three broad groupings (see Appendix for details). The fixed exchange-rate regime included pegs to single currencies and pegs to the SDR or other baskets. The second regime allowed for some flexibility in the nominal rate and included countries in the European Monetary System (EMS), pegs which are adjusted frequently and managed floating rates. The third regime consisted of freely floating exchange rates. In practice, these three groups represent relative degrees of flexibility rather than precise distinctions. For expositional purposes, we refer to these regimes as fixed, managed and floating exchange-rate regimes.

Our broad classification scheme is similar to that adopted by the IMF. It differs from Mussa by differentiating between managed and floating exchange-rate regimes which he had combined into a single group. The other difference is due to our focus on effective RERs. For example, as already mentioned in our classification scheme, Austria was fixed to a major currency until 1990 and so it was included in the fixed exchange-rate regime throughout this period.

The accuracy of our classification scheme depends partly on the accuracy of the information that member countries submit to the IMF. The IMF attempts to assess the accuracy of this information and on occasion shifts countries into nominal exchange-rate regime categories accordingly.

Ideally, the classification of the nominal exchange-rate regime should be based on a country having in place a broad set of policies which are consistent with a given regime. However, on occasion, a country will claim to be in a given regime but run policies which are inconsistent with this regime. For example, a country with very high inflation that attempts to be in a fixed regime, but does nothing other than switch to a fixed regime in order to control inflation, is unlikely to be able to hold a fixed peg for an extended period. Therefore, such a country will display a highly variable RER (and nominal exchange rate) simply because the peg has to be devalued in large discrete steps in order to prevent sustained appreciations of the RER. In practical terms, such a country is not genuinely in a fixed exchange-rate regime. We addressed this problem by examining a subset of countries with low and stable inflation and stable growth rates that were most likely to be accurately classified in terms of their exchange-rate regimes.[15]


Quarterly data was available for two additional countries, New Zealand and Papua New Guinea. [14]

Also, for reasons that we do not fully understand, many developing countries (which also have high inflation) report their regimes as floating when they clearly are not. [15]